Title :
Analytic model for MPEG-4 and H.263 encoded video traces
Author_Institution :
Dept. of Comput. Eng., KFUPM, Dhahran, Saudi Arabia
Abstract :
Summary form only given. We present the analysis of statistical distributions of 1-hour long publicly available empirical samples of MPEG-4 and H.263 encoded video traces. Each video has been encoded to different quality levels. The results of our analysis show that: i) although the autocorrelation of the studied traces asserts long-range dependence, however, the distribution of the same traces is not heavy-tailed; ii) for different quality levels, the traces belong to the same statistical distribution; iii) most of the studied traces show long-tail behavior which can be modeled using lognormal distribution function.
Keywords :
correlation theory; encoding; fractals; image sampling; log normal distribution; video coding; H.263 encoded video traces; LLCD; MPEG-4 encoded video traces; analytic model; empirical samples; heavy-tailed distribution; lognormal distribution; long-tail behavior; quality level; statistical distribution analysis; Autocorrelation; Distributed computing; Distribution functions; MPEG 4 Standard; Statistical distributions;
Conference_Titel :
Computer Systems and Applications, 2003. Book of Abstracts. ACS/IEEE International Conference on
Conference_Location :
Tunis, Tunisia
Print_ISBN :
0-7803-7983-7
DOI :
10.1109/AICCSA.2003.1227566